|
Abstract:
|
The attribution of the causes of trends in the amount of nutrients transported by streams is often limited to correlation, qualitative reasoning, or simply references to the work of others. This paper considers efforts to improve causal attribution of water-quality changes using the Red River of the North (Manitoba, Minnesota, North Dakota, and South Dakota) and the Chesapeake Bay (Delaware, District of Columbia, Maryland, Pennsylvania, New York, Virginia, and West Virginia) watersheds as test cases. These watersheds were selected because of strong regional and international interest in the reduction of nutrient transport, long-term water-quality monitoring, and the availability of ancillary geospatial data with the potential to explain changes in loads. Multiple working hypotheses and structural equation modeling methods are investigated to test causal hypotheses related to the potential explanations of nutrient changes. Methods for estimating missing ancillary data, and water-quality related challenges to structural equation modeling (including skewed data and scaling issues) are also explored. The resulting models provide quantitative attribution of water-quality trends.
|
